Fibroblast growth factor reverses the bacterial retardation of wound contraction

Fibroblast Growth Factor Reverses the Bacterial Retardation of Wound Contraction Peter Hayward, MB, BS, FRACS, James Hokanson, PhD, John Heggers, PhD, Galwston,Texas,John Fiddes, PhD, Corine Khngbeil, PhD, Mountain View,California, Mare Goeger, MS, Martin Robson, MD, Galveston,Texas

Chronic granulating wounds were established in rats by excising burns inoculated with Escherichia coli. Recombinant human basic fibroblast growth factor was applied at dosages of 1, 10, and 100 &cm2 to the wounds of three groups of 20 animals on days 5,9, 12, 15, and 18 after injury. The rate of wound closure was compared with that of similarly wounded animals treated with saliue vehicle alone. High levels of bacteria caused significant retardation of wound contraction. The addition of basic fibroblast growth factor at the 100 &cm2 dosage level markedly improved the rate of wound closure whereas inert vehicles applied alone were ineffective. Since bacterial counts did not decrease in the basic fibroblast growth factor treated wounds, basic fibroblast growth factor was not inherently bactericidal. Histologic examination of the wounds treated with basic fibroblast growth factor showed increased cellularity with increased nmnhers of fibroblasts and round cells. These results suggest basic fibroblast growth factor can overcome the defect in healing created by bacterial infection, and this peptide may have efficacy in the management of the contaminated wound.

From the Division of Plastic Surgery and the Wound Healing Lahoratory (PH, JH, JH, MG, MR), University of Texas Medical Branch and Shriners Burns Institute, Galveaton, Texas, and California Biotechnology, Inc. (JF, CK), Mountain View, California. Requests for reprints should be addressed to Martin Rohson, MD, Division of Plastic Surgery, University of Texas Medical Branch, Galveston, Texas 77550. Manuscript submitted December 10,1990, and accepted in revised form April 2,199 1. 288

THE AMERICAN JOURNAL OF SURGERY

T

he identification and purification of peptide growth factors and related cytokines has expanded our knowledge of wound healing. While it is clear that growth factors are involved in the various processes of healing, their precise individual roles in normal healing processes and the clinical indications for which growth factor therapy might be useful are both yet to be determined. Since being identified [I], the fibroblast growth fattors have been shown to have numerous in vivo and in vitro effects, suggesting roles as wound healing agents. Of the two commonly available forms, basic fibroblast growth factor is generally considered to be the more potent mitogen in vitro and is the most widely studied. It has been shown to have an important role in both angiogenesis and the fibroblastic response to injury. In particular, basic fibroblast growth factor has been shown to be rnitogenie and chemotactic to endothelial cells [2-51 and to promote capillary budding [a. Basic fibroblast growth factor is also chemotactic for fibroblasts, stimulating their entry into the wound, the multiplication of resident fibroblasts, and the enhancement of matrix and collagenase production [ 7-14. Finally, basic fibroblast growth factor has recently been shown to stimulate keratinocyte replication [II]. Topical application of basic fibroblast growth factor to uncompromised animals results in a modest acceleration of healing in incisional [12] wounds. In normal animals, anti-basic fibroblast growth factor antibodies retard granulation tissue formation in a polyvinyl alcohol sponge model [ 131. In vivo effectiveness has been demonstrated in incisional and polyvinyl alcohol sponge methods in diabetic rats [14,15] and in full-thickness wounds in diabetic, obese, and steroid-treated mice [Id]. Infection is a potent cause of retarded healing [ 17-211. The infected open wound is a clinical problem that is currently treated with topical antimicrobials. However, recent evidence suggests that topical antimicrobials show toxicity in vitro to cells involved in wound healing, particularly fibroblasts [22] and keratinocytes [23]. In the clinical management of contaminated wounds, agents that restore healing processes without causing cellular damage would be desirable. This study was designed to evaluate the role of basic fibroblast growth factor in a chronically infected granulating wound model. MATERIALS AND METHODS

Chronic granulating wounds were prepared in our laboratory as previously described [24-26J. Male Sprague Dawley rats weighing 300 to 350 g were used after being acclimatized for a week in our facility prior to use. Using intraperitoneal pentobarbital to anesthetize the

VOLUME 163 MARCH 1992

FIBROBLAST GROWTH FACTOR IN WOUND CONTRACTION

rats (35 mg/kg), the rat dorsum was shaved and depilated. A full-thickness dorsal burn measuring 30 cm2 was created by immersion in boiling water. Infected groups were seeded with 108Escherichia coli (ATCC #25922, American Type Culture Collection, Rockville, MD) after the rats had been allowed to cool for 15 minutes. Bacteria were obtained from fresh 18-hour broth cultures, and inoculm size was confirmed by back plating. All animals were resuscitated with 7 mL subcutaneous Ringer’s solution. Animals were individually caged and given food and water ad Zibitum.Uninfected animals were kept in a physically separate facility. All experiments were conducted in accordance with the Animal Care and Use Committee guidelines of the University of Texas Medical Branch. Five days after burning, the eschar was excised from anesthetized animals resulting in a chronic granulating wound. The histology of this wound has been previously character&d and compared with a human chronic granulating wound [26J. The animals were divided into five groups treated as follows: Group 1: Twenty animals that had not been inoculated acted as uninfected controls and received no topical treatment. Group 2: Forty rats that had been infected with lo8 colony forming units of E. coli had their wounds excised and were treated with the same topical vehicle as that used for basic fibroblast growth factor dilutions (vehicle controls) on days 5,9,12,15, and 18. The saline vehicle used was formulated as 50 mm01 sodium acetate at pH 5.5 adjusted to isotonicity with normal saline. Group 3: Twenty-fwe infected animals had their wounds treated with basic flbroblast growth factor 1 rg/ cm2 topically on days 5, 9, 12, 15, and 18. Group 4: Twenty-five infected animals had their wounds treated with basic Bbroblast growth factor 10 pg/cm2 topically on days 5, 9, 12, 15, and 18. Group 5: Twenty-five infected animals had their THE AMERICAN

wounds treated with basic libroblast growth factor 100 pg/cm2 topically on days 5, 9, 12, 15, and 18. The basic fibroblast growth factor used was the human 154 amino acid form produced by recombinant expression in E. coli [27] and was provided by P. Shadle and L. Foster at California Biotechnology, Inc. (Mountain View, CA). All wounds were left exposed. Any dried exudate that formed was atraumatically removed to allow direct application of the test substance to the wound surface. Wounds were biopsied for quantitative bacteriology at debridement and at 13 and 19 days after escharectomy to exclude superinfection and to confirm bacterial levels in infected animals. The wound surface was cleaned with 70% isopropyl alcohol prior to biopsy to exclude surface contamination. Biopsies were weighed, homogenized, and back plated onto nonselective media. Bacterial counts were made after incubation of 48 hours and expressed as colony forming units (CFU) per gram of tissue [28]. Every 48 hours the outline of the wounds was traced onto acetate sheets, and area calculations were performed using computerized digital planimetry (Sigma&an, Jande1 Scientific, Corte Modera, CA). Care was taken only to record the perimeter of the wound that represented the advancing full-thickness margin rather than the edge of any advancing epithelium. All animals were weighed on a weekly basis. Five animals from each of the infected groups (2,3,4, and 5) were killed by pentobarbital overdose for histologic examination on day 19, after all treatment doses had been given. Three full-thickness, transverse strips of granulation tissue were then harvested from the proximal, middle, and caudal ends of the wound and fmed in 10% buffered formalin. Transverse sections (5 pm) were taken from each specimen and stained with hematoxylin and eosin. The thickness of the granulation tissue was estimated with an ocular micrometer at low power. High powered fields were examined immediately below the

JOURNAL

OF SURGERY

VOLUME

163

MARCH

1992

289

TABLE I

QuantitativeBacteriology Results for Wound Biopsies Collected 13 and 19 Days After Eschar Removal*

I Group

Day13

Day 19

(Uninfected controls) (Infected, vehicle controls) (bFGF I pg/cm2) (bFGF 10 bg/cm2) 5 (bFGF 100 pg/cmq



1 2 3 4

bFGF

= basic 6broMast

*Results

are mean

growth

counts

factor.

for treatment

groups

expressed

in colony

forming

units per gram of bcdy tissue.

1.200, INFECTED ;z

1.000

bFGF 9

CONTFtOL5

NONINFECTED

0.600

3a

o-o

CONTROLS.

1 0Dw/cm2

-

.

*---A

bFGF

10~/cm20-•o

bFGF

l&cm2

A-A

0.600 B 5

RESULTS Quantitative bacteriology: Biopsy of all noninfected

0.400

g d LL 0.200

0.000 0

5

10

15

DAY5

20

POST

25

30

35

CREATlON

40

45

50

55

60

OF GRANULATING

65

70

75

60

WOUND

Flgwe 2. Percentage area of wound remaining cpen plotted

against timeafter creationof ganulatlng woundand commence mentoftre&wntfcrallbeatmentgwps.

TABLE II Hlstometric Data for Wound Blopsles Taken at 19 Days* Group Fibroblastcount Round cells Capillaries Cellularity+ *Counts tCellularity

represent

2 31.3 2 20.9” 3.1 5 2.9 + mean

3 3.1 2.0 0.6 0.2

+ standard

37.1 34.1 3.5 4.0

k 2 f 2

4 4.1 3.4 0.3 0.3

40.6 34.9 2.7 4.4

5

f + 2 t

3.9 4.2 0.3 0.3

41.7 29.3 3.3 3.6

k k 2 +

3.6 3.2 0.4 0.3

deviation.

rated on a scale from 1 to 5 with highly cellular

biopsies

rated 5.

superficial inflammatory layer of the granulation tissue (F’lgure 1). From each strip of granulation tissue, five adjacent high powered fields were photographed and coded. Enlarged prints of these exposures were then used for hiitometric analysis in a blinded fashion. Fibroblasts, “round” cells (macrophages, lymphocytes, and neutro phils), and capillaries were counted. In addition, the cellularity of each section was graded on a scale of 1 (reduced cell counts) to 5 (highly cellular). The remaining animals were killed by pentobarbital overdose when the wound had healed or when it had 290

reached a point where the wound was less than 10% of its original area. Measurement of very small wounds by manual tracing introduces significant systematic error, and we have found that wounds followed past this point may remain static for prolonged periods of time. Statistical analysis: Serial area measurements were plotted against time. For each animal’s data, a Gompertz equation was fitted (typical r2 = 0.85) [24]. Using this curve, the wound “half-life” was estimated. Comparison between groups was performed using life table analysis and the Wilcoxon rank test. These statistical analyses were performed using the SAS and BMDP (BMDP Statistical Software, Inc., Los Angeles, CA) packages on a personal computer. Cell counts for the different treatment groups were pooled and analyzed using a one-way analysis of variance. Post hoc analyses of differences between groups were carried out using Bonferroni’s modification of the ttest with p 10.05 considered significant. These latter analyses were performed using the NCSS (Number Cruncher Statistical System, Kaysville, UT) statistical package.

THE AMERICAN JOURNAL OF SURGERY

groups showed less than lo* organisms/gram of tissue on day 13 and 19. Infected animals showed consistent tissue levels of bacteria greater than lo5 E. coli CFU/gram of tissue both in basic fibroblast growth factor treated and non-basic fibroblast growth factor treated wounds (Table I).

Body weighm There was an equivalent gain in body weight among all groups during the period of study, and there was no significant variation among the groups. Wound area: Figure 2 shows plots of the mean wound area versus time for all treatment groups. Infected animals demonstrate markedly different decay curves than those of uninfected rats indicating a retardation of wound contraction. Life table analysis shows that this retardation of closure is significant (p
VOLUME 163 MARCH 1992

FIBBOBLAST GROWTH FACTOR IN WOUND CONTRACTION

fibroblast counts were statistically different only in a comparison of the infected controls (Figure 3A) and those animals treated with the highest concentration of basic fibroblast growth factor (F’igure 3B). Sections from all basic flbroblast growth factor treated animals were more celhrlar and had higher numbers of round cells in the sections examined. COMMENTS These results indicate that tissue contamination with viable bacteria significantly retards the rate of contraction of an open wound in the rat. This is in agreement with other studies for both acutely [29] and chronically [24,25] contaminated wounds. Basic fibroblast growth factor can reverse the retardation of wound closure in this granulating wound model but only when administered at 100 pg/cm2 5 times over 18days(Figure 2). AdministraTHE AMERICAN

tion of basic libroblast growth factor at 1 and 10 pg/cm2 on the same dosage schedule had no effect. Treatment with basic fibroblast growth factor enhanced the cellularity of the wound, particularly its round cell infiltrate and, in the case of the highest concentration, the number of fibroblasts. These findings are in accordance with other histologic studies of the effect of this factor in mice [JO]. In another experiment (unpublished data), a single dose of 100 rg/cm2 applied to this model at the time of wound debridement did not enhance contraction of the chronic granulating wound. This is in contradistinction to other data reported that the use of a single dose of basic libroblast growth factor has had a pronounced effect on wound closure. In the acute contaminated wound, Stenberg et al [29] demonstrated that a single dose of basic fibroblast growth factor could overcome the inhibition to contraction caused by bacteria. Similarly, Klingbeil et al

JOURNAL

OF SURGERY

VOLUME

163

MARCH

1992

291

[16] found that single doses of 1 to 10 pg/cm2 were effective in the closure of full-thickness wounds in diabetic, obese, and steroid-treated mice. The apparent failure of a single large (3 mg) dose to have a significant effect upon this model whereas repeated doses were successful must be carefully interpreted. If basic fibroblast growth factor is maximally active after the initial “lag” phase of wound closure when contraction enters its rapid exponential phase, then repeated pulsed doses could have greater efficacy as seen in this study. The finding that a high dose was required to enhance contraction in this study differs from the conclusions of experiments with noninfected animal wounds. High tissue levels of bacteria or bacterial products such as prots ases may bind or inactivate the growth factor necessitating the high levels of basic fibroblast growth factor needed to produce an effect in this model. Nonspecific protein effects do not seem to be involved since topical treatment with high concentrations of human serum albumin alone does not produce beneficial effects on the rate of contraction in this model (data not shown). The mechanism by which basic libroblast growth factor accelerates wound closure in this model is unclear. Certainly, bacterial infection affects all phases of the wound healing response [22]. It is not readily conceivable how basic fibroblast growth factor could reverse bacterial distortion of prostaglandin, complement, or free radical levels seen in the inflammatory phase of wound healing. Similarly, it is unlikely to reverse the platelet and leukocyte dysfunction seen in wound infection. Bacterial infeu tion also affects the angiogenesis and synthesis of the extracellular matrix that occurs in the proliferative phase of wound repair. Macrophages stimulated by bacterial products, particularly endotoxin, stimulate enhanced angiogenesis [31]. Topical application of endotoxin produces increased granulation tissue mass [32]. Although basic fibroblast growth factor has been shown to be a potent angiogenic agent [6,33,34], it did not produce a change in capillary density or granulation tissue thickness in this study. The cellular effects that were seen were primarily an increase in the number of fibroblasts and round cells. In the wound itself, there is a balance between the production and destruction of extracellular matrix. By increasing the number of fibroblasts, basic fibroblast growth factor may enhance production of ground substance; however, this was not measured in this study. In contrast, there is clear evidence that high tissue levels of bacteria reduce the amount of collagen deposition and increase the rate of collagen breakdown. Increased collagenolytic activity may be due to the action of bacterial proteases [35,36] or bacterial stimulation of macrophages, which, in turn, elucidate collagen cleaving enzymes [37]. Bacteria have been reported to increase fibrinolysis by the secretion of plasminogen activators and other proteases [38]. Basic fibroblast growth factor may act to restore production of the elements of the extracellular matrix since it is known to enhance production of fibre nectin and proteoglycans [ 121. Considering the multiple in uitro qualities of basic fibroblast growth factor, this factor could reverse bacteri292

THE AMERICAN JOURNAL OF SURGERY

al “poisoning” of the wound at a number of different levels in the sequence of wound repair. Further studies will be necessary not only to define the mechanisms involved but also the appropriate time/dose relationships for topical treatment in this model.

REFERENCES 1. Gaspodarowicz D. Localizationof a fibroblastgrowth factor and its effect alone and with hydrouxtisone on 3T3 cell growth. Nature 1974; 249: 123-7. 2. Moscatelli D, Presta M, Ritkin DB. Purification of a factor from human placenta that stimulates capillary endothelial cell protease production, DNA synthesis and migration. Prcc Nat1 Acad Sci USA 1986; 83: 2091-5. 3. Connolly DT, Stoddard BL, Harakas NK, Feder J. Human fibroblast-derived growth factor is a mitogen and chemoattractant for endothelial cells. Biochem Biophys Res Commun 1987; 144: 705-12. 4. Presta M, Moscatelli D, Joseph-Silverstein J, Rifkin DB. Purification from hepatoma cell line of a basic fibroblast growth factorlike molecule that stimulates capillary endothelial cell plasminogen activator production, DNA synthesis and migration. Mol Cell Biol 1986; 6: 4060-6. 5. Montesano R, Vassalli JD, Baird A, Guillemin R, Orci L. Basic fibroblast growth factor induces angiogenesis in vitro. Proc Nat1 Acad Sci USA 1986; 83: 7297-301. 6. Folkman J, Klagsbrun M. Angiogenic factors. Science 1987; 235: 442-7. 7. Gospodarowicz D, Massoglia S, Cheng J, Lui GM, Bohlen P. Isolation of pituitary fibroblast growth factor by fast protein liquid chromatography (FPLC). Partial chemical and biological characterization. J Cell Physiol 1985; 122: 323-93. 8. Vlodavsky I, Johnson LK, Greenburg G, Gospodarowicz D. Vascular endothelial cells maintained in the absence of fibroblast growth factor undergo structural and functional alterations that are incompatible with their in uiuodifferentiated properties. J Cell Biol 1979; 83: 468-86. 9. Davidson JM, Klagsbrun M, Hill KE, ef al. Accelerated wound repair, cell proliferation and collagen accumulation are produced by a cartilage-derived growth factor. J Cell Biol 1985; 100: 1219-27. 10. Buckley-Sturrcck A, Woodward SC, Senior RM, Griffin GL, Klagsbrun M, Davidson JM. Differential stimulation of collage nase and chemotactic activity in fibroblasts derived from rat wound repair tissue and human skin by growth factors. J Cell Physioll989; 138: 70-89. Il. O’Keefe RJ, Chin ML, Payne RE. Stimulation of growth of keratinocytes by basic fibroblast growth factor. J Invest Dermatol 1988; 90: 767-9. 12. McGee GS, Davidson JM, Buckley A, et al. Recombinant basic fibroblast growth factor accelerates wound healing. J Surg Res 1988; 45: 145-53. 13. Broadley KN, Aquino AM, Woodward SC, et al. Monospecific antibodies implicate basic fibroblast growth factor in normal wound repair. Lab Invest 1989; 61: 571-5. 14. Phillips LG, Geldner P, Brou J, Dobbins S, Hokanson J, Robson MC. Correction of diabetic incisional wound impairment with basic fibroblast growth factor. Surg Forum 1990; 41: 602-3. 15. Broadley KN, Aquino AM, Hicks B, et nl. Growth factors basic fibroblast growth factor and TGF beta accelerate the rate of wound repair in normal and in diabetic rats. Int J Tissue React 1988; 6: 345-53. 16. KlingbeiI CK, Cesar LB, Fiddes JC. Basic fibroblast growth factor: effects on tissue repair in models of impaired wound healing.

VOLUME 163 MARCH 1992

FIBROBLAST GROWTH FACTOR IN WOUND CONTRACTION

In: Barbul A, Caldwell M, Eaglestein W, er al, editors. Clinical and experimental approaches to dermal and epidermal repair: normal and chronic wounds. New York: Alan R. Lii, 1991: 443-50. 17. Bucknall TE. The effect of local infection upon wound healing: an experimental study. Br J Surg 1980; 67: 851-5. 18. deHaan BB, Ellis H, Wilks M. The role of infection on wound healing. Surg Gynecol Obstet 1974; 138: 693-700. 19. Dom M, Russwurm R. Tierexperimentelle untersuchungen zur beeinflussung der wundheilung durch Candidaalbicum.Arch Dermat01 Res 1976; 256: 205-12. 20. Smith M, Enquist IF. A quantitative study of impaired healing resulting from infection. Surg Gynecol Obstet 1967; 125: 965-73. 21. Robson MC, Stenberg BD, Heggers JD. Wound healing alterations caused by bacteria. Clin Plast Surg 1990, 17: 485-92. 22. McCauley RL, Linares HA, Pelligrini V, Hemdon DN, Rob son MC, Heggers JP. In uitro toxicity of topical antimicrobial agents to human fibroblasts. J Surg Res 1989; 46: 267-74. 23. McCauley RL, Poole B, Heggers JP, Robson MC, Hemdon DN. Differential in vitro toxicity of topical antimicrobial agents to human keratinocytea. Proceedings of the 22nd Annual Meeting of the American Bum Association. 1990; 2. 24. Hayward PG, Robson MC. Animal models of wound contraction. In: Barbul A, Caldwell M, Eaglestein W, et al, editors. Clinical and experimental approaches to dermal and epidermal repair: normal and chronic wounds. New York: Alan R. Liss, 1991: 30112. 25. Robson MC, Krizek TJ. The effect of human amniotic membranes on the bacterial population of infected rat bums. Ann Surg 1973; 177: 144-9. 26. Robson MC, Edstrom LE, Krizek TJ, Groskin MG. The eflicacy of systemic antibiotics in the treatment of granulating wounds. J Surg Res 1984; 16: 299-306. 27. Thompson SA, Protter AA, Bitting L, Fiddes JC, Abraham JA. Cloning, recombinant expression and characterization of basic fibroblast growth factor. In: Barnes D, Mather J, Sato G, editors.

Methods in enzymology. New York Academic Press, 1990: In press. 28. Heggers JP. Variations on a theme. In: Heggers JP, Robson MC, editors. Quantitative bacteriology: its role in the armamentariurn of the surgeon. Boca Raton: CRC Press, 1991. 29. Stenberg BD, Phillips LG, Hokanson JA, Heggers JP, Robson MC. Effect of basic libroblast growth factor on the inhibition of contraction caused by bacterial contamination. Surg Forum 1989; 40: 629-31. 30. Greenhalgh DG, Sprugel KH, Murray MJ, Ross R. PDGF and FGF stimulate wound healing in the genetically diabetic mouse. Am J Path01 1990; 136: 1235-46. 31. Hunt TK, Knighton DR, Thakral KK, Goodson WA, Andrews WJ. Studies on inflammation and wound healing: angiogenesis and collagen synthesis stimulated in vivo by resident and activated wound macrophages. Surgery 1984; 96: 48-54. 32. Kanta J, Novotny J, Pumprila P, Cerna E, Bartos F. The effect of endotoxin on granulation tissue formation in rats. Exp Path01 1981; 20: 230-2. 33. Gospodarowicz D, Bialecki H, Thakral TK. The angiogenic activity of fibroblast and epidermal growth factor. Exp Eye Res 1989; 28: 501-14. 34. Gospodarowicz D, Baird A, Cheng J, Lui GM, Esch F, Bohlen P. Isolation of fibroblast growth factor from bovine adrenal gland: physicochemical and biological characterization. Endocrinology 1986; 118: 82-90. 35. Dunphy J. The cut gut. Am J Surg 1970; 119: l-8. 36. Niinikoski J, Grislis G, Hunt TK. Respiratory gas tensions and collagen in infected wounds. Ann Surg 1972; 175: 588-93. 37. Wahl LM, Wahl SM, Mergenhagen SE, Martin GR. Collagen production by endotoxin-activated macrophages. Proc Nat1 Acad Sci USA 1974; 71: 3598-601. 38. Perry AW, Sutkin HS, Gottlieb LD, Sadelman WK, Krizek TJ. Skin graft survival-the bacterial answer. Ann Plast Surg 1989; 22: 479-83.

THE AMERICAN JOURNAL OF SURGERY

VOLUME 163 MARCH 1992

293